1. APPENDIX
TABLES AND FIGURES
Table A : Summary Statistics
Table B : Skewness/ Kurtosis Normality Test:
joint
Variable Obs Pr(Skewness) Pr(Kurtosis) adj chi2(2) Prob>chi2
log(CO2) 188 0.0033 0.2213 9.1 0.0106
joint
Variable Obs Pr(Skewness) Pr(Kurtosis) adj chi2(2) Prob>chi2
Log(gdp) 180 0.4903 0.0000 21.90 0.0000
joint
Variable Obs Pr(Skewness) Pr(Kurtosis) adj chi2(2) Prob>chi2
Log(upop.) 206 0.0000 0.7922 17.27 0.0002
joint
Variable Obs Pr(Skewness) Pr(Kurtosis) adj chi2(2) Prob>chi2
Log(vden.) 57 0.6559 0.5616 0.55 0.7603
joint
Variable Obs Pr(Skewness) Pr(Kurtosis) adj chi2(2) Prob>chi2
[Log(vden.)]2
57 0.0193 0.9348 5.31 0.0703
Variable Observations Mean Std. Dev. Min Max
CO2 188 5.423822 8.050425 0.0126951 69.15906
Log(CO2) 188 0.6543592 1.712385 -4.366537 4.236409
GDP 180 6678.001 9977.794 86.45013 49996.1
Log(GDP) 180 7.680186 1.60252 4.459568 10.8197
Urban Population 206 55.96155 24.5921 9.72 100
Log(Urban Population) 206 3.90114 0.5396325 2.274186 4.60517
Vehicle Density 57 50.21053 56.43287 3 254
Log(Vehicle Density) 57 3.407637 1.021376 1.098612 5.537334
[Log(Vehicle Density)]2
57 12.6369 7.210388 1.206949 30.66207

2. Regression Results:
Answer 1
Table 1.1: Regression of logCO2 on logupopulation & logvehicledensity
Source SS Df MS Number of obs = 56
Model 39.2060945 2 19.6030473 F( 2, 53) = 30.73
Residual 33.8075327 53 0.637877975 Prob > F = 0
Total 73.0136272 55 1.32752049 R-squared = 0.537
Adj R-squared = 0.5195
Root MSE = 0.79867
logCO2 Coeff. Std.Err. t P>|t| [95% conf. Interval]
Logupop. 1.69305 0.3757207 4.51 0 .93945 2.44665
Logvden. 0.3391488 0.1278357 2.65 0.011 .0827429 .5955547
_cons -6.763162 1.366831 -4.95 0 -9.504678 -4.021647
Table 1.2: Mean of error terms
Mean Std.Err. [95% Conf. Interval]
e -3.04e-09 0.1047687 -.2099611 .2099611
Table 1.3(a): Correlation matrix of error terms & explanatory variables
e Logupop. Logvden.
e 1
Logupop. 0 1
Logvden. 0 0.5755 1
Table 1.3(b): VIF
Variable VIF 1/VIF
Logupop. 1.5 0.668764
Logvden. 1.5 0.668764
Mean 1.5
Table 1.4: Heteroskedascity Test
Breusch-Pagan / cook-Weisberg test for heteroskedasticity
Ho: Constant variance
Variables: fitted values of logCO2
chi2(1) = 0.4
Prob > chi2 = 0.4923
*Logupop. = log(urbanpopulation) , logvden. = logvehicledensity

3. Table 1.5: Misspecification Test
Table 1.6: Normality of error terms
Jarque-Bera normality test: 0.7748 Chi(2) .6788
Jarque-Bera test for Ho: normality
Answer 2
Table 2.1: Regression of logCO2 on log(urbanpopulation), log(vehicledensity) & rich
Source SS df MS Number of obs = 56
Model 49.3314345 4 12.3328586 F( 3, 52) = 26.56
Residual 23.6821927 51 0.464356719 Prob > F = 0
Total 73.0136272 55 1.32752049 R-squared = 0.6756
Adj R-squared = 0.6502
Root MSE = 0.68144
logCO2 coeff. Std.Err. t P>|t| [95% Conf. Interval]
Logupop. 1.120989 0.3562969 3.15 0.003 .4056934 1.836285
Logvden. 2.4028 0.5250698 4.58 0.000 1.348679 3.456922
Rich 0.6663922 0.2399022 2.78 0.008 .1847685 1.148016
(Logvden.)2
-0.3036461 0.0733061 -4.14 0.000 -0.4508142 -.156478
_Cons -7.8182 1.426956 -5.48 0.000 -10.68294 -4.953466
Table 2.2: Mean of error terms
Mean Std.Err. [95% Conf. Interval]
e 1.31e-10 0.876871 -0.1757288 0.1757288
Table 2.3(a): correlation matrix of error terms & explanatory variables
e Logupop. Logvden. Rich (Logvden.)2
e 1
Logupop. -0.0000 1
Logvden. -0.0000 0.5755 1
Rich -0.0000 0.5726 0.4970 1
(Logvden.)2
-0.0000 0.5481 0.9844 0.5030 1
Ramsey RESET test using powers of the fitted values of logCO2
Ho: model has no omitted variables
F(3, 50) = 2.13
Prob > F = 0.1084

4. Table 2.3(b): VIF
Variable VIF 1/VIF
Logupop. 1.85 0.541369
Logvden. 34.65 0.028857
Rich 1.63 0.614730
(Logvden.)2
33.62 0.029747
Mean 17.94
Table 2.4: Heteroskedascity Test
Breusch-Pagan / Cook-Weisberg test for heteroskedasticity
H0: Constant variance
Variables: fitted values of logCO2
chi2(1) = 1.50
Prob > chi2 = 0.2209
Table 2.5: Misspecification Test
Ramsey RESET test using powers of the fitted values of logCO2
Ho: model has no omitted variables
F (3, 48) = 1.88
Prob > F = 0.1460
Table 2.6: Normality of errors
Jarque-Bera normality test: 2.485 Chi(2) 0.2886
Jarque-Bera test for Ho: normality

5. Answer 3
Table 3.1: Regression of logCO2 on log (urbanpopulation), log (vehicle density), dummies for Asia,
North America, South America, Oceania and Europe
Source SS df MS Number of obs = 56
Model 49.859338 7 7.12276257 F(7, 48) = 14.77
Residual 23.1542892 48 0.482381025 Prob > F = 0
Total 73.0136272 55 1.32752049 R-squared = 0.6829
Adj R-squared = 0.6366
Root MSE = 0.69454
logCO2 Coeff. Std.Err. t P>|t| [95% Conf. Interval]
Logupop. 1.366379 0.3838719 3.56 0.001 0.594553 2.138206
Logvden 0.2830142 0.1232192 2.3 0.026 0.0352654 0.530763
As 0.6984331 0.3815024 1.83 0.073 -0.0686288 1.465495
Na 0.4135816 0.4331655 0.95 0.344 -0.4573561 1.284519
Sa -0.3027211 0.4843232 -0.63 0.535 -1.276518 0.671076
Eu 1.048195 0.3373656 3.11 0.003 0.3698761 1.726514
Oc 1.349596 0.6169087 2.19 0.034 0.1092183 2.589975
_Cons -5.916017 1.290214 -4.59 0 -8.510166 -3.321868
Table 3.2: Mean of error terms
Mean Std. Err. [95% Conf. Interval]
e -2.18e-09 0.0867042 -0.1737592 0.1737592
Table 3.3(a): Correlation matrix of error terms & explanatory variables
e logupop logvden As Na Sa Oc Eu
e 1
Logupop 0 1
Logvden 0 0.5755 1
As 0 0.1218 0.288 1
Na 0 0.0021 0.0548 -0.1548 1
Sa 0 0.0641 -0.2177 -0.1371 -0.0868 1
Oc 0 0.1651 -0.0581 -0.0951 -0.0603 -0.0534 1
Eu 0 0.1452 0.0646 -0.4771 -0.3021 -0.2676 -0.1857 1
As = Asia, Na = North America, Sa = South America, Oc = Oceania, Eu = Europe

6. Table 3.3(b): VIF
Variable VIF 1/VIF
Eu 3.3 0.30312
As 2.67 0.374955
Logupop 2.06 0.484488
Logvden 1.84 0.544344
Sa 1.81 0.553662
Na 1.77 0.564586
Oc 1.52 0.657223
Mean 2.14
Table 3.4: Heteroskedascity
Table 3.5: Misspecification Test
Table 3.6: Normality of errors
Breusch-Pagan / Cook-Weisberg test for heteroskedasticity
Ho: Constant variance
Variables: fitted values of logCO2
chi2(1) = 0.50
Prob > chi2 = 0.4801
Ramsey RESET test using powers of the fitted values of logCO2
Ho: model has no omitted variables
F (3, 45) = 2.40
Prob > F = 0.0803
Jarque-Bera normality test: .4714 Chi (2) .79
Jarque-Bera test for Ho: normality

7. Answer 4
Table 4.1: Regression logCO2, GDP, log(urbanpopulation), log(vehicledensity), Asia, South America,
North America, Europe, Oceania, North America1, South America1, Europeu1, Oceania1, Asia1
Source SS df MS Number of obs = 55
Model 57.6125472 13 4.4317344 F( 13, 41) = 11.8
Residual 15.4006539 41 0.375625705 Prob > F = 0
Total 73.0132011 54 1.35209632 R-squared = 0.7891
Adj R-squared = 0.7222
Root MSE = 0.61288
loglogCO2 Coeff. Std.Err. t P>|t| [95% Conf. Interval]
Logupop. 0.9552699 0.4046581 2.36 0.023 .1380463 1.772494
Logvden. 0.2605639 0.1275214 2.04 0.047 .0030292 .5180985
Asia 1.843853 0.5034198 3.66 0.001 .8271758 2.860529
Europe 1.954247 0.502562 3.89 0 .9393023 2.969191
NA 0.5495919 0.6296615 0.87 0.388 -.7220354 1.821219
Oceania 0.3393984 2.241271 0.15 0.88 -4.186941 4.865737
SA 0.760279 0.8204543 0.93 0.36 -.8966621 2.41722
As1 -0.0004321 0.0001546 -2.8 0.008 -.0007442 -.00012
Eu1 -0.0004167 0.0001592 -2.62 0.012 -.0007383 -.0000952
Na1 -0.0003575 0.0001609 -2.22 0.032 -.0006825 -.0000325
Oc1 -0.0003127 0.0001983 -1.58 0.122 -.0007131 .0000877
Sa1 -0.000426 0.0003179 -1.34 0.188 -.0010681 .000216
Gdp 0.0004284 0.0001609 2.66 0.011 .0001036 .0007533
_Cons -5.178985 1.368565 -3.78 0 -7.942857 -2.415112
Table 4.2: Mean of error terms
Mean Std.Err. [95% Conf. Interval]
e 9.57e-10 0.0720098 -.1443709 .1443709
As1 = Asia*Gdp, Na1 = North America*Gdp, Sa1 = South America*Gdp, Oc1 = Oceania*Gdp, Eu1 = Europe*Gdp

8. Table 4.3(a): Correlation matrix of error terms & explanatory variables
E Logco
2
Gdp Log-
upop
Log-
vden
As Sa Na Eu Oc Na1 Sa1 Eu1 Oc1 As
1
e 1.0
0
logco2 0.4
6
1
gdp 0.0
0
0.61 1
logupo
p
0.0
0
0.69 0.57 1
logvde
n
0.0
0
0.60 0.46 0.59 1
As 0.0
0
0.12 0.02 0.12 0.30 1
Sa 0.0
0
-0.27 -0.21 0.06 -0.22 -0.14 1
Na 0.0
0
-0.07 0.01 0.04 0.01 -0.14 -0.08 1
Eu 0.0
0
0.37 0.25 0.14 0.08 -0.49 -0.28 -0.28 1
Oc 0.0
0
0.16 0.12 0.16 -0.06 -0.10 -0.05 -0.05 -0.19 1
Na1 0.0
0
0.16 0.25 0.09 0.04 -0.09 -0.05 0.63 -0.17 -0.03 1
Sa1 0.0
0
-0.22 -0.18 0.09 -0.16 -0.12 0.89 -0.07 -0.24 -0.05 -0.04 1
Eu1 0.0
0
0.39 0.75 0.30 0.16 -0.30 -0.17 -0.17 0.62 -0.12 -0.11 -0.15 1
Oc1 0.0
0
0.17 0.13 0.16 -0.07 -0.10 -0.05 -0.05 -0.19 0.98 -0.03 -0.05 -0.11 1
As1 0.0
0
0.25 0.28 0.38 0.59 0.71 -0.10 -0.10 -0.35 -0.07 -0.06 -0.09 -0.22 -0.07 1

9. Table 4.3(b): VIF
Variable VIF 1/VIF
Gdp 564.93 0.00177
Eu1 520.21 0.001922
As1 156.21 0.006402
Na1 91.75 0.010899
Oc1 74.67 0.013392
Oc 25.77 0.038799
Eu 9.24 0.108197
Sa 6.65 0.150445
Sa1 6.01 0.166453
As 5.94 0.168427
Na 3.91 0.255431
Logupop. 2.93 0.341874
Logvden. 2.5 0.399485
Mean 113.13
Table 4.4: Heteroskedascity
Breusch-Pagan / Cook-Weisberg test for heteroskedasticity
Ho: Constant variance
Variables: fitted values of logCO2
chi2(1) = 0.01
Prob > chi2 = 0.9078
Table 4.5: Misspecification Test
Ramsey RESET test using powers of the fitted values of logCO2
Ho: model has no omitted variables
F(3, 38) = 2.08
Prob > F = 0.1185
Table 4.6: Normality of errors
Jarque-Bera normality test: 2.886 Chi(2) .2362
Jarque-Bera test for Ho: normality

10. FIGURES
Figure 1: Plot of predicted errors from the regression of log(CO2) on log(urban population) and
log(vehicle density)
Figure 2: Plot of predicted errors from the regression of log(CO2) on rich dummy, log(urban
population), log(vehicle density) and [log(vehicle density)]2
0
.2.4.6
Density
-2 -1 0 1 2
Residuals
0
.2.4.6
Density
-1 0 1 2
Residuals

11. Figure 3: Plot of predicted errors from the regression of log(CO2) on continent dummies, log(urban
population) and log(vehicle density)
Figure 4: Plot of predicted errors from the regression of log(CO2) on GDP per capita, continent
dummies, interaction dummies, log(urban population) and log(vehicle density)
0
.2.4.6
Density
-2 -1 0 1 2
Residuals
0
.2.4.6.8
1
Density
-1 -.5 0 .5 1 1.5
Residuals